China’s Merger Control Policy: An Initial Review of Patterns and Implications Xinzhu Zhang JXUEF, SHUEF and CASS Tokyo, February, 2009

Outline Brief description of competition policy regime Developments of enforcement Preliminary analysis of five milestone cases Patterns and Implications Conclusions

Background and motivations After AML took effect in August 2008, enforcement has become the central focus. However, NDRC and SAIC just start to enforce non-merger antitrust rules and have not announced any significant cases yet. As the agency in charge of merger review, MOFCOM (or more precisely, the Anti-monopoly Bureau) has reviewed under the new pre-merger notification regulation more than 60 cases with relevant turnovers surpassing the notification thresholds. For the time being, most cases (more than 90%) were decided in Phase I.

But five cases were challenged - InBev/Anheuser Busch, Coca-Cola/Huiyuan, Mitsubishi Rayon/Lucite, Pfizer/Wyeth, and Panasonic/Sanyo. InBev/Anheuser Busch, Mitsubishi Rayon/Lucite, Pfizer/Wyeth, and Panasonic/Sanyo were approved with restrictions. Coca-Cola/Huiyuan was blocked, the first one since China installed the new merger control regime. This presentation will analyze the patterns and implications of China’s merger control policy based on preliminary economic analysis of these significant cases.

These cases received worldwide attentions. Regulations and guidelines are newly released or are still being prepared; the merger control agency is struggling to find consistent methods to implement merger control rules; limited information is available. So uncertainties exist for AML enforcement. Since information is very limited, it is hard to do economic analysis, both theoretically and empirically. And any analysis and interpretation of current merger control policy might be speculative. However, it is still valuable to analyze these significant cases to help understand China’s merger control policy.

Preliminary analysis of the milestone cases I. The InBev/Anheuser Busch Case On 13 July 2008, Belgium based beer giant InBev and the US based brewery Anheuser Busch announced InBev’s acquiring of Anheuser Busch for US$49.91 billion. The case was approved within 30 days limit of Phase I after MOFCOM found the transaction will NOT eliminate or restrict competition in China’s beer market.

However, since pre-merger Anheuser-Busch had a 27% stake in Tsingtao Brewery, China’s second largest beer producer, and InBev had a 29% stake in Zhujiang Brewery, China’s fourth largest beer producer, MOFCOM imposed restrictions: Post-merger InBev should not increase its stakes in Tsingtao Brewery and Zhujiang Brewery from pre- merger levels; InBev should not acquire any stakes in China Resources Snow Breweries or Beijing Yanjing Brewery, the largest and third largest beer producers in China.

Pre-merger both InBev and Anheuser have significant presence in China’s beer market as producers and equity investors. No indication of how relevant market was defined even though terms like market (beer market) and market share were mentioned. In particular, the relevant geographical was implicitly defined as national but it was unclear how the geography market was dealt with. Information on market shares and concentration level was not released. The case was treated as a horizontal merger. MOFCOM worried about potential collusive effect between the merged party and its rivals post-merger.

Indeed, the main competition concern came from the significant interests that InBev/AB had in local giant beer producers - the second biggest stake holder in both Tsingtao Brewery and Zhujiang Brewery. MOFCOM was concerned that the proposed acquisition of AB by InBev might facilitate tacit collusion and reduce competition if InBev/AB further increased its shareholding in Tsingtao Brewery and Zhujiang Brewery or if the merged party would acquire interests in the other two largest beer producers, i.e. China Resources Snow Breweries or Beijing Yanjing Brewery, China’s first and the third largest beer producers.

Presumption on coordinated effect was mainly based on shareholding in rivals but no further analysis on the structural factors that may facilitate or prevent collusion. Interestingly, as a global merger case, the same merger had also been cleared in United States with DOJ imposing divestiture of InBev’s sales force in some geographical markets to ensure effective competition. The decision was largely consistent with international practice - it is neither more restrictive nor more relaxed than other jurisdictions.

II. The Coca-Cola/Huiyuan Case On September 2, 2009, US soft drinks giant Coca-Cola offered to buy the Chinese leading juice maker Huiyuan Juice Group for US$ 2.4 billion. MOFCOM decided to enter Phase II after 30 days the case was accepted given the large scale and considerable influence of this concentration. Exactly 90 days later, MOFCOM decided to block the proposed merger.

MOFCOM ruled that Coca-Cola would be capable of extending its dominant position in the carbonated soft drinks (CSDs) market to the fruit drink market post merger and foreclose potential competitors to enter the fruit juice market. Coca-Cola might remarkably enhance its market power post merger by controlling two well-known brands (portfolio effect), which also constitute entry barriers to potential competitors to enter the fruit juice market.

Coca-Cola and Huiyuan overlapped in fruit juice market and Huiyuan was dominant in high concentrated fruit juice. It seemed that when defining the relevant product market, MOFCOM was concerned mainly by whether fruit juice with different concentration levels should be treated as within the same relevant market. The fruit juice market was defined as the relevant product market. Definition of relevant product market was claimed to be based on substitution analysis but no details were revealed on how the substitution pattern, particularly substitution between the carbonated soft drink and fruit juice, was analyzed.

MOFCOM revealed no information on market shares and concentration level in the fruit juice market. Marker shares in fruit juice market were different with different sources. For example, one source shows that top three fruit juice producers, Uni-President, Coca-Cola and Huiyuan had market shares of 18.69%, 15.04%, 13.95%, respectively, in Thus post merger Coca- Cola/Huiyuan would have a marker share of and become the largest fruit juice producer. In fact, all sources seemed to indicate that post merger Coca- Cola/Huiyuan’s market share is less than 30%. It was unclear how MOFCOM choose between different sources.

Perhaps because market share is low, MOFCOM did not presume horizontal effects even though it is a concentrated market and the proposed merger would lead to significant increase of HHI ( △ HHI >390). MOFCOM also calculated that Coca-Cola has a market share of 60.6% in the CSDs market and thus it was presumed to have significant market power in this market. MOFCOM presumed further that the merged party would leverage its market power to the juice market and use tying, bundling or discriminatory pricing strategies to foreclose competitors in the fruit juice market and thus hurt consumers.

MOFCOM mentioned the strong brand effect as entry barrier to the fruit juice market to satisfy its burden of proof. MOFCOM did not prove that the vertical constraints would indeed be anti-competitive and the merging parties have both the incentive and the ability to bundle and tie the products in order to leverage market power from one market into another, i.e. to foreclose competitors to enter the juice market.

It seemed that MOFCOM has crafted narrow market definition to address competition concerns with foreclosure effect but did not pay attention to the fact that it entails higher standard of burden of proof. In fact, narrow market definition is not necessary for antitrust authorities to address anticompetitive concerns. Presumption of unilateral effect might still be possible if post-merger there was significant loss of local competition between the two brands. No indication of how the rebuttal from the merging party was considered.

III. Mitsubishi Rayon/Lucite On September 11, 2008, Japanese chemical giant Mitsubishi Rayon Co. announced its acquisition of UK plastics maker Lucite International Group for US$1.6 billion. After a four-month probe, MOFCOM expressed its concerns that the proposed merger could hurt competition given that pre-merger two merging parties have a combined market share of 64% for methyl methacralate (MMA) in China.

MOFCOM imposed divestiture plus code of conduct: Lucite China should divest 50% of its annual MMA production capacity for a one-time sale to one or several non-related third-party buyers which have the right to buy MMA at a price just covering production cost for five years. Lucite should operate independently from the MMA monomer business operations of Mitsubishi Rayon China until the completion of capacity divestiture. Both Mitsubishi Rayon and Lucite are restricted on further acquisitions and new plant construction in mainland China.

MOFCOM defined MMA as the relevant product market and the relevant geographical market as national. MOFCOM calculated post-merger market share of the merged party as 64% and thus presumed post-merger Mitsubishi Rayon/Lucite has significant market power. Indeed, it depended on this high market share to presume anti-competition effects. An interesting issue related to calculation of market share was a German producer was due to production in late 2009 but this new capacity was not taken into account. If so, market share would be less than 40%.

MOFCOM presumed two competition concerns. One was horizontal effect, that is the merged party might eliminate or restrict competition unilaterally due to its market power in the MMA market. MOFCOM was also concerned by vertical effect -the merged party might leverage its market power in the MMA market to the downstream markets and foreclose competition in downstream markets. Again MOFCOM provided no evidence to proof both the unilateral effect and the foreclosure effect and no indication of the rebuttal process.

IV. Pfizer/Wyeth On June 9, 2009, US pharmaceutical giant Pfizer filed to MOFCOM for the proposed acquisition of another US pharmaceutical producer Wyeth. On September 29, 2009, MOFCOM approved this proposed transaction with restrictions. Structural remedies were imposed in which post-merger Pfizer should divest its production assets in mainland China of swine vaccines for porcine enzootic pneumonia under the brands of Respisure and Respisure One.

Relevant product market was defined as swine vaccines for porcine enzootic pneumonia and relevant geographical market was defined as national. The merger party has combined market share of 49.4% (Pfizer 38% and Wyeth 11.4%), which is far exceeding that of the second largest producer (8.35%). Pre-merger the market is highly concentrated (HHI is 2182) and the proposed merger would lead to significant increase of HHI ( △ HHI would be 336). MOFCOM presumed unilateral effect based on these market structure parameters.

However, it was unclear how the substitute pattern between relevant products of different brands were analyzed. And there is no indication of whether market shares were reflective of not only the first choice but also the second choice of consumers of the products produced by the merged party. MOFCOM did not indicate to what extent there is significant price effects. While market structure factors were used to presume competition harm, there is no information on the rebuttal process.

V. Panasonic/Sanyo On January 21, 2009, Japanese giant conglomerate Panasonic filed to MOFCOM for its proposed acquisition of another Japanese large manufacture Sanyo. On October 30, 2009, MOFCOM approved this transaction but with a series of structural remedies. The first remedy was Sanyo’s production facility in Japan for rechargeable coin-shape batteries based on lithium should be divested and sold to the third party.

Secondly, Sanyo’s production facilities in Japan for portable rechargeable nickel-metal hydride batteries was divested and sold to the third party. In addition, Sanyo’s production facility in Suzhou, China or Panasonic’s production facility in Wuxi, China was divested and sold to the third party. Finally, Panasonic’s production assets for primary cylindrical lithium batteries in Japan and Hunan, China were divested and sold to the third party.

Rechargeable coin-shape batteries, portable rechargeable nickel-metal hydride batteries and primary cylindrical lithium batteries were defined as the relevant product markets. Relevant geographic markets were defined as international. The merger party has post merger market share of 61.6% (merger of the largest and second largest producers) in rechargeable coin-shape batteries markets. MOCOM based on high market shares to presume unilateral effect in this market. Buyer market power was dismissed.

The merger party has post merger market share of 46.3% in portable rechargeable nickel-metal hydride batteries market. Again, MOFCOM based on this magnitude of combined market share to presume unilateral effect. The merger party has combined post merger market share of 77% in the primary cylindrical lithium batteries market, which allows MOFCOM to presume unilateral effect in this market.

Relevant product market is implicitly defined based on functionalities of different products. No information is available on how substitution pattern was analyzed. Indeed, it is unclear to what extent market shares are reflective of not only first choice as well as the second choice of consumers of relevant products. MOFCOM did not indicate the magnitude of price effects.

Patterns and Implications Market definition Market definition has played pivotal role but it seemed that it was done mainly based on analysis of functionality of products. No indication of SSNIP analysis being used (even no simply use of critical loss analysis). And no diversion ratio analysis in the case of differentiated markets. It may therefore cast doubt on the precise of market definition. For example, in the Coca-Cola/Huiyuan case, it is key to analyze the substitutability between carbonated soft drink and fruit juice products and it would be hard to draw the product market line without detailed quantitative analysis of substitution patterns.

In some cases the relevant geographical market concern may not be addressed adequately. In markets like beer or juice products with important distribution channel restrictions and transport costs, relevant geographical market is likely to be regional. If relevant geographical market was not defined correctly as regional, anti-competition concerns could only be addressed at the national level. For example, the InBev/AB case may only raise anti-competition concerns in some regional markets and thus remedies should be designed accordingly.

Using market shares to presume market power and competition harm Until recently, MOFCOM tended to rely on high market shares to presume market power (Art. 16 of AML) and competition harm, that is to establish prima facie case. While market shares provide only incomplete information about market power, this practice is useful and effective given its enforcement capability. In fact, even in jurisdictions such as the US and EU with rich antitrust enforcement experiences, market share information also plays an important role.

But presumptions of market power and competition harm based on market shares should be subject to rebuttal, which implies significant part of burden of proof should shift to the merging party. Therefore, a due procedure is important allowing proper rebuttal process. It is a challenge for administrative enforcement system like China where such mechanisms are still being built up.

MOFCOM tended to rely on narrow market definition or high market shares and dominant position to presume unilateral effect but it may not be necessary based on economic theory. For one thing, market definition is likely to be dubious in differentiated markets. For another, what matter for unilateral excise of market power is loss of local competition or the extent to which the proposed merger can internalize loss of sales due to price increases and thus change the incentives of the merged party to raise prices. In other words, market shares may not reflect precisely the substitute pattern between products in the relevant market.

Presuming non-horizontal competition harms The MOFCOM has dealt with both horizontal (coordinated effect in InBev/AB and unilateral effects in Mitsubishi Rayon/Lucite, Pfizer/Wyeth, and Panasonic/Sanyo) as well as non-horizontal mergers, but it presumed too often foreclosure effects (Coca- Cola/Huiyuan and Mitsubishi Rayon/Lucite) without realizing the necessary and heavy burden of proof for such presumptions of competition harms. International experiences seems to suggest that antitrust agencies are usually more prudent in presuming non-horizontal competition harms.

Not much attention has been paid to the fact that higher standard of proof is required for presuming non- horizontal competition harms. Market foreclosure is an legitimate anti-competition concern and thus due care should be taken. But higher standard of proof is required – the agency need not only to proof anti- competition conducts are likely to arise but also the merged party has the incentive and ability to do so.

Imposing behavioral remedies While structural remedies were most often used, in some cases behavior remedies were imposed, which might be a problem for enforcement as behavior remedies are usually more difficult to enforce. For example, in the Mitsubishi Rayon/Lucite case price regulation was imposed. Some remedies may be itself anticompetitive in some cases. For example, in the Mitsubishi Rayon/Lucite case the merged party was required not to increase its capacity.

Independence of decision making Because of institutional design, the independence of antitrust scrutiny was doubted. Some consider politics and industrial policy concerns have exerted great influence on decision. But others believe decisions so far are based solely on professional antitrust analysis. For instance, some argued that because of change of capital markets, Coco-Cola benefited from rejection of the Coca-Cola/Huiyuan case. In addition, this case decision is consistent the Coco-Cola/Berri case in Australia. It is too early to judge.

Conclusions MOFCOM has obviously consulted the decisions of other jurisdictions but in the meantime addressed local competition concerns. Indeed, so far decisions are largely consistent with international practice. MOFCOM is obviously developing its capability to deal with cases very promptly, e.g. reasoning is becoming more solid with increasingly more economic analysis, more and more information was released and so on.

Some avenues to improve Information disclosure is one aspect that China’s antitrust authorities should and can improve in a short- term period. More generally, building a due process with a proper rebuttal process is particularly important. Economic analysis should be more extensively adopted to provide more evidence.

Thank you!